How to Make an AI Zombie Love Story Video With Two Photos

2026-10-16T06:55:00.000Z
VideoAny AI Zombie Couple studio showing separate Photo 1 human and Photo 2 zombie identity upload slots
VideoAny's two identity-reference slots, captured October 16, 2026. Photo 1 supplies the human; Photo 2 supplies the zombie. The picture on the right is an external scene reference.

Categories: AI Creation Studio, Halloween, Photo to Video

Tags: AI zombie love story, AI zombie couple, Halloween AI video, photo to video, VideoAny effects

The AI zombie love story trend is a short fictional reunion: one person stays human, the other is styled as a gentle, non-graphic zombie, they recognize each other and share a quiet embrace. You do not need costumes, makeup, or a filmed set. You need two clear portraits and a tool that already knows the scene.

On VideoAny, use AI Zombie Couple. The current effect settings, as of October 16, 2026, are:

SettingCurrent value
InputTwo JPG or PNG photos, up to 20 MB each
RolesPhoto 1: human; Photo 2: zombie identity reference
ModelSeedance 2.5
Output5 seconds, 720p, vertical 9:16
AudioSilent; no generated music, voices, or sound effects
Cost568 credits per generation

Check the effect page for the latest settings and credit cost before generating. This guide covers photo choice, upload roles, generation, review, and downloading your clip.

What the effect creates

Photo 1 is the human identity reference. Photo 2 is the identity reference for the person AI is asked to style as a pale, subtle zombie. These are character references, not the first and last frames of the video. The fixed prompt asks for two distinct people in a moonlit reunion and gentle embrace. Facial details, clothing, and poses may change, so judge the whole clip rather than the opening frame alone.

The intended mood is tender Halloween storytelling. The prompt excludes blood, gore, wounds, and aggressive actions. These are scene instructions, so watch the full clip to check that the result fits your audience.

External dramatic human-and-zombie reunion reference showing two adults embracing on an empty street
External reunion reference from zombie-couple.com, with more distressed zombie styling. VideoAny's prompt aims for a gentler appearance. This image is not a VideoAny-generated result.

Step 1: Pick two clear individual photos

Use one person per image. Favor sharp, well-lit faces with eyes, nose, and mouth visible. Matching framing (both head-and-shoulders, or both waist-up) helps identity stay stable.

Skip group shots, sunglasses, heavy beauty filters, masks, and faces half-hidden by hair or objects. You do not need a couple selfie, and you do not need a real zombie-makeup photo for image 2. A normal portrait is enough for the zombie styling.

Only use photos you have permission to use, and make sure both people are fine appearing in a fictional romantic Halloween scene.

Step 2: Upload human first, zombie second

Open AI Zombie Couple.

  1. Upload one clear JPG or PNG in Photo 1 · Human identity reference.
  2. Upload the other person's JPG or PNG in Photo 2 · Zombie identity reference.
  3. Keep each file at or below 20 MB and check both previews before you spend credits.

The upload slots tell the prompt which person has each role. If the previews are reversed, correct the files before generating. This page uses a fixed scene prompt; there is no custom prompt to fill in.

Step 3: Check credits, then generate one test

Sign in and check the displayed cost and your balance. The setting as of October 16, 2026 is 568 credits for one 5-second, 720p clip.

Select Generate once and allow several minutes for processing. Five seconds is the video length, not the generation time. Keep the page open while it processes. If progress cannot be confirmed, check your generation history before starting another task.

Start with one take. Keep it for comparison and replace only one reference photo at a time if you need to improve a face; each new generation uses credits.

Step 4: Review identity, embrace, and crop

Watch the whole clip once, then scrub the recognition moment and the embrace.

External human-and-zombie embrace reference showing the placement of faces, arms, and hands during a hug
Use the embrace as a reference for checking faces and overlapping arms. Source: zombie-couple.com. The appearance and motion of a VideoAny clip will vary.

Check for:

  • Face drift on either person
  • The wrong person receiving zombie styling, despite the intended upload roles
  • Merged hands or bodies during the hug
  • Faces cut off in a vertical 9:16 crop

If only one face drifts, replace that portrait with a sharper one and keep the other file. If the roles are wrong, compare the previews with the slot labels first. Correct reversed uploads; if they were already correct, the model may have mixed the roles. Try clearer, visually distinct portraits while keeping Photo 1 as the intended human and Photo 2 as the intended zombie. Keep the take where both people stay recognizable through the embrace.

When you are happy with the result, select Download video. Save the file before preparing captions or adding audio in your editor or publishing app.

Before you share

Preview at the size your audience will see on TikTok, Reels, or Shorts. Keep captions away from faces and hands. Label the clip as AI-generated fiction. The downloaded clip is silent, with no generated music, voices, or sound effects. Add audio afterward and use music you have the right to share.

A short caption that supports the story without over-explaining works well, for example "I'd still know you" or "One soft reunion." Avoid presenting the scene as a real event.

Troubleshooting before another generation

  • The task is still running or the page reports a progress error: check your generation history before trying again. A missing preview does not prove that the submission failed.
  • The credit status is unclear: check the task and your balance, then contact support if the status remains uncertain. Do not submit repeated jobs while resolving it.
  • One face changes or the two identities merge: use separate, sharp photos with similar framing, unobstructed faces, and one person in each image. Replace the weaker reference first instead of changing both at once.
  • The embrace hides a face or crops it out: try waist-up portraits with space around the head and shoulders. The scene is generated in vertical 9:16, so the original crop is an identity reference rather than a guaranteed final composition.
  • The wrong person becomes the zombie: check the Photo 1 and Photo 2 labels. If the uploads were already correct, preserve the intended roles and try clearer references; reversing correct inputs is not a reliable fix.

FAQ

Which photo becomes the zombie?
The prompt assigns zombie styling to Photo 2 and keeps Photo 1 human. Check the finished clip because AI can mix the roles.

Can I use one couple photo?
Use two separate images with one person each. That matches the live two-photo workflow.

Do I need a prompt or a reference video?
No. The effect page already includes the scene direction. You supply the two identity photos.

Is it free?
The current setting as of October 16, 2026 is 568 credits per generation. Check the effect page and your balance before you start.

Will it look gory?
The effect aims for pale, non-graphic styling and a tender mood. Results can still vary, so review before posting.

Make your clip

Choose two portraits, open AI Zombie Couple, and place the intended human in Photo 1 and the intended zombie in Photo 2. Check the current cost, generate one take, then review both identities and the embrace before downloading.

For a one-photo Halloween scene, follow the AI broom riding video guide or the AI ghost prank video guide. You can also explore AI Pumpkin Head.